Simulation of realistic audio-visual environments for audiological research.
Stefan Kerber and Bernhard U. Seeber
MRC Institute of Hearing Research, University Park, Science Road, Nottingham. 
NG7 2RD.

Free-field presentation of sounds from various loudspeaker positions is a 
prerequisite for assessment of spatial hearing abilities and for research on 
binaural hearing. Presentation of sounds from loudspeakers in the free-field has 
several advantages over headphone presentation: Since people sit in a free-field 
they are listening with their own ears and thus with their own, by definition 
correct, head-related transfer functions. This is particularly important when 
testing hearing-impaired subjects, as they can listen with hearing devices in the 
way they are used and adapted to. Cross-comparisons between normal-hearing and 
hearing-impaired subjects can thus be done reliably. Additionally, head-movements 
are possible and alter the sound in the familiar, correct way. 

    The Simulated Open Field Environment (SOFE, Hafter and Seeber, 2004, 
Kirszenstein, 1984) allows simulation of realistic audio-visual scenes in an 
anechoic chamber. Sound-fields consisting of multiple sources and their 
reflections can be simulated with the mirror-image method. Simulated sources and 
reflections are played from loudspeakers that are placed in the anechoic chamber 
at or near the direction of the simulated source. Because sounds are played in an 
anechoic chamber, interaction of the simulated reflections with the existing 
reflections in the anechoic test room is kept at a minimum. In addition to the 
sound playback, visual scenes can be projected synchronised to the sound. Subject 
responses can be evaluated to interactively modify the audio and the visual scene. 

     The poster presents a refined version of the SOFE which was recently built at 
the MRC Institute of Hearing Research. Small, round loudspeakers were designed to 
cover a frequency range of 200-12000 Hz with a single, broadband chassis while 
allowing sound playback at up to 90dB SPL at the subject’s head. The reduced phase 
distortion from using a broadband chassis is particularly useful for spatial 
hearing research while the high output level is crucial for work with unaided 
hearing-impaired patients. Individual loudspeaker equalisation assures a 
frequency-independent response and calibrated levels across all speakers. 36 
loudspeakers are placed horizontally at ear height and a further 12 are mounted at 
elevated positions. Additionally, visual projection was extended from one frontal 
to three screens, widening the possibilities for audio-visual experiments 
tremendously.

Acknowledgements
This work was supported by the MRC. 

References
Hafter E., Seeber B., 2004. Redesign of the simulated open-field environment and 
it's application in audiological research. In P.A. Santi, editor, Abstracts of the 
28th annual midwinter meeting. Assoc. Res. Otolaryngol, p. 339.
Kirszenstein, J., 1984. An image source computer model for room acoustics analysis 
and electroacoustic simulation. Applied Acoustics 17: 275-290.